The Hannibal Route Question of 218 Bc: a Forensic Exercise Relative to Historical Archaeology

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by ZENODO

DOI: 10.5281/zenodo.1323871

THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE RELATIVE TO HISTORICAL ARCHAEOLOGY

William C. Mahaney

Quaternary Surveys, 26 Thornhill Ave., Thornhill, Ontario, Canada, L4J 1J4, and Department of Geography, York University, 4700 Keele St., Toronto, Ontario, Canada, M3J 1P3 e-mail: [email protected]

Received: 29/06/2018 Accepted: 05/08/2018

ABSTRACT Following a long and protracted survey of all targeted approach routes, cols of passage and exfiltration pathways projected to have been followed by Hannibal and his generals when they crossed into Italy in 218 BC, the physical evidence points to the Col de la Traversette, first identified by Sir Gavin de Beer in the 1960‘s. The first attempts to identify the route out of a dozen possible transits, focused not only on historical interpretations using the evolution of place names but on physical evidence possibly resident in hearths, alluvial terraces and rock rubble masses along the various approach routes. The primary argument following ten years of investigating every approach route from the Col Agnel in the south to the Col Mt. Cenis in the north, was that if the only blocking rockfall described by Polybius was present below the Traversette col, then mires or fans on either side in France and Italy might contain a record of Hannibal‘s passage. Beyond their key positions as water sources and foraging areas, the French mire and coalescing alluvial fan sediment in the upper Po, might carry evidence of the ecologic disturbance that could be radiocarbon dated to the Hannibal time line of 2168 cal yr BP or 218 BC. It is this long quest to unlock the Hannibalic invasion route that opened up key areas for historical archaeological exploration. This forensic cross/discipline exercise might serve to highlight a valuable method useful in solving other elusive ancient historical archaeological problems.

Keywords: Hannibalic invasion route; forensic analysis of mire sediment; Po plains; Polybius; environmental 46 W.C. MAHANEY

1. INTRODUCTION the three routes, historic speculation aside, might be closest to the actual route traveled by the Punic Ar- The question of the route followed by Hannibal my. and his army in 218 BC, 2168 cal yr BP (BP=before Despite the thinking of some workers that the present, 1950 AD) has rankled and irked historians landscape has changed greatly since the time of and classicists for millennia (Freshfield, 1886, 1899; Hannibal‘s invasion, it is clear from environmental Dodge, 1891; de Montholon, 1905; Wilkinson, 1911; summaries published by Sodhi et al. (2006), Ma- Walbank, 1956, 1990; Brown, 1963; de Beer, 1956, haney (2008a) and Mahaney et al. (2008a, b, c; 2016a, 1967, 1969, 1974; Seibert, 1993) amongst others. No 2018a) that the landscape we see today, despite less than seven books were written about the inva- many slight modifications by stream erosion and sion in Hannibal‘s time (Proctor, 1971), the last avalanche activity over the last two millennia, is the known of which (by Silenus, Hannibal‘s historiog- same as that seen by Hannibal and his army when rapher) was destroyed by the great fire that con- they crossed into Italia in 218 BC. The landscape was sumed the library at Alexandria in 300 AD. Since set in place during the Late Pleistocene (the last ice then a number of workers have posited possible age as it were), the final touches taking place during routes (Fig. 1), basically dependent upon previous the Bølling and Allerød interstades that marked the sources including the texts of Livy (trans. de Sélin- first warming time since the retreat of ice, followed court, 1965) and Polybius (trans. Paton, 1922; Scott- by a cosmic airburst of 12.8 ka (12,800 yrs) that fired Kilvert, 1979), but few have visited the sites in ques- the surface leaving a carbonized record in the land tion making their target cols all based on second- surface now resident in paleosols (ancient soils) and hand information. While some have tried to correlate rock rinds (Mahaney and Keiser, 2013; Mahaney et descriptions in ancient texts to the actual landscape al., 2013, 2014, 2016a, 2017d, 2018b). This airburst and flooding times of rivers (de Beer, 1969), none event, termed the black mat is also defined as the have coupled the present environment with major YDB-Younger Dryas Boundary-that marks the tran- topographic features in the ancient texts, let alone sition into a cold period called the Younger Dryas searched hearths, likely campgrounds, possible rock- (YD, 12.8-~11.5 ka), a term resulting from earlier pol- fall impediments to the exfiltration of the mountains. len work that identified variable pollen concentra- It is the purpose of this summary to outline the tions of the cold loving plant (Dryas octapetala) in thinking that went into trying to elicit an environ- many alpine regions. The distribution of this plant mental matrix from the ancient texts, one that could increased during the Younger Dryas event, a glacial be tested against the present landscape, a forensic resurgence that resulted from a cold time, similar to exercise if you will (Mahaney, 2004, 2008a; Mahaney a nuclear winter, projected to have been caused by et al., 2008a, b, c). the black mat airburst. Assessing the Hannibalic Route one is faced with The black mat event is known to have affected countless authorities arguing for one of the three North, Middle and South America, Greenland, west- principal routes (Fig. 1) over the other and most key ern Europe, Central Asia, (Wolbach, 2018a, 2018b), authorities (Proctor, 1971; Lazenby, 1998; Lancel, and is thought to have reached Antarctica (Mahaney 1999) opting for the northern route principally be- et al., 2018c). Its effect on the Western Alps, particu- cause of lower elevation and ease of access. Only de larly the area around the Mt. Viso massif was pro- Beer (1967, 1969, 1974), Prevas (1998), and Bagnall found, burning the land surface and creating the (1999) favor the southern route principally because largest accumulation of soot recorded so far in West- times of flooding, place name matches, and view ern Europe (Allen West, personal communication, onto the Po plains closely align with Polybius and 2016). Portions of bedrock and much exposed glacial Livy. The middle route outlined by Connolly (1981) sediment were carbonized and melted by tempera- pitches indecision into Hannibal‘s thinking that once tures topping 2000 oC. The event presumably melted along the Isère and near Grenoble he would have all or most glacial vestiges of the ice age, to be short- shifted priority to finding the Col de Genèvre, the ly replaced by a resurgence of ice forming during the most unlikely deviation of travel plans. To enter into ensuing nuclear winter which produced revived this historic mix, a daunting task as it were, required positive glacial mass balances for more than a mil- a test of environmental parameters elicited from the lennium. ancient texts of Livy and Polybius, forming a matrix of interconnected variables to determine which of

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 47

Fig. 1. Invasion routes as proposed by several classicists. Figure reprinted after Mahaney et al. (2017a). Hannibal’s Rhône crossing point is not known with certainty but is probably near Arles south of Avignon. The ‘Island’, Hannibal’s resupply point is not known exactly but is probably near present-day Orange. Major tribes – Insubres and Ligures, allied with Hannibal are shown within their nomadic areas; the Taurini refused overtures from Hannibal and were deci- mated in late 218 BC. The Boii tribe, allied with Hannibal, occupied an area to the east near present-day Bologna. The YD event lasted for ~1.3 kyr, ending by ~11.5 of elevation mentioned in ancient texts, the study ka when glaciers worldwide went into recession, soon morphed into a forensic test requiring the use nearly disappearing in some cases by the Holo- of methods from diverse disciplines to answer ques- cene/Pleistocene boundary (~10 ka). By this time the tions related to camping places, foraging areas, de- landscape through which Hannibal passed had been posit impediments, river crossings and other envi- set in place, all further change attributed to minor ronmental targets as they might have affected time amounts of air-influxed material into soils and pale- motion analysis of the Punic Army. osols and mass wasted sediment additions to valley Polybius provides us with specific day-indicators side locations (e.g. the younger rockfall sheet identi- of the invasion which allow for a total time travel of fied by Polybius) in local valleys). Losses of sedi- five months for the march from New Carthage to the ment accrued in valleys by stream erosion and de- Po plains, with four days from the Rhône crossing to velopment of Holocene terraces only slightly modi- the island, ten days from the island to the beginning fying the landscape in the area of the Col de la of the ascent of the mountains, followed by fifteen Traversette. The impediments to exfiltration faced by days to reach the col. The assumption is that the fif- Hannibal are the very same as exist today on both teen days are included in the five months but Polyb- the French and Italian sides of the Western Alps. It is ius is not certain about this. We do not know what within this matrix of bedrock and deposits that the Polybius meant by ‗ascent of the Alps‘ as this quote search for environmental evidence related to the in- may mean ascent of the Dauphiné Alps or ascent vasion began in 2002. Starting as a means of match- following the debacle in the Combe de Queyras ing topographical descriptions and relative changes (Mahaney and Tricart, 2008). Taking the Rhône

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 48 W.C. MAHANEY

crossing to the island to be approximately 40 km, the the land surface or in section below the surface, must next leg to the Col de Grîmone a total of 180 km and occupy a tangential position in the reconstruction on allowing for negotiations and resupply efforts with both the French and Italian sides of the Alps. Be- local tribes at Orange to take 2-3 days, it is likely this cause most historians writing over centuries about might consume two weeks. If this is correct then the the Punic Wars recounted historical documentation fifteen days to reach the col would include 105 km- only with little regard to time/motion analysis of the Col de Grîmone to Durance basin, 5 km Durance to Punic Army, and with scant attention paid to envi- Guil catchment, 50 km Guil to Traversette, a total of ronmental variables recounted here, it is easy to see 160 km and would entail significant time to offset why belief structures aligned with different groups the Durance crossing difficulties, battle in the Combe favored one route over another. This, combined with de Queyras, and reorganization in the upper Guil a lack of ground truthing of routes tied to the ancient probably taking six to eight days. The distance from literature, led to embedded historical hypotheses the col to the Po plains is 30 km, the last ten km tak- lacking any evidence providing an outcome with en up traveling through a steep gorge. little chance of finding artifacts that might support a Of the three routes identified in Fig. 1, both the particular route. northern and southern routes are approximately the Despite other route hypotheses expounded upon same distance, only the terrain difference causing a by many historians, including mention of routes slower rate on the southern leg. The middle route closer to the Mediterranean coast and others further favored by Connolly (Fig. 1) is longer by about 50 afield in Switzerland, such as the Great St. Bernard km and would have brought the Punic army to with- Pass one has to consider Hannibal‘s prime strategic in the territory of the Allobroges for a longer period objective pointed out by Hart (1967)---Linkage with that would have entailed many delays due to en- Gallic tribes in cis-alpine Gaul—the Ligurian, Boii, gagements relative to the southern route. Putting and Insubre tribes with whom Hannibal had estab- time distance together for the full route, Polybius lished contact prior to the invasion. Travelling too gives the distance as 9000 stadia (1597 km). During close to the Mediterranean Coast would allow the the first century BC, Strabo calculated 8.5 stadia to Romans to maintain contact along the way and trav- the Roman mile which equates to one stadium equal elling north of the Isère River would add to to 177.8 m. The Roman stadium is 0.1778 km, some- transport time, thus causing delays. Considering that what shorter than an Olympic stade which is 0.184 Hannibal had reasonable information from agents km or 185 m. Distances were summarized by de Beer sent to contact Gallic tribes in 219, returning to New (1969) who assumed Strabo‘s calculations equaled Carthage (Polybius, III, 34) in spring 218 BC, and as Polybius‘ distances converted to km‘s as cited above. expressed by Polybius, it is reasonable to assume he Of all the environmental parameters mentioned in had knowledge of all major passes leading out of the ancient texts of Polybius and Livy, including: trans-alpine Gaul and that his objective col was the nighttime attack on a col along the approach, snow- Col de Genèvre at ~2000m asl. Because the Gauls line elevations, frozen ground (possibly permafrost), were shadowing him following departure of low- long defile below the ascendant col of passage, land Gallic cavalry sent from Orange as guides (Po- camping ground near the summit, hearths, view lybius III, 50), the Allobroges must have assumed across the Po river plains, rockfall blockade during Hannibal‘s intended crossing target, and realizing exfiltration to the Po river, regrouping area on the they could easily attack and defeat him with his cav- lee side, the one key element to identifying the route alry ineffective in the narrow confines of the pass, is the large, robust two-tier rockfall. All cols of pas- they closed up on his rearguard (Polybius III, 52). sage have rockfall on the lee side of the Alps but the Thus, with the Allobroges behind and ahead at his only pass with a large rubble sheet on the Po river col target, either Hannibal or his second-inside, one which could have impeded exfiltration, lies command, Maharbal, must have deviated into the below the Col de la Traversette (~3000 m asl). As Guil River and the gorge (Combe de Queyras) just discussed by Mahaney et al. (2008c, 2010b, 2010c), beyond the Durance/Guil confluence (Fig. 1) as a this is the key element in the route puzzle that has means of bypassing the Allobroges. Aside from the mystified historians and classicists for over two mil- rockfall on the lee side of the range, no other ap- lennia, and the one geomorphic feature as described proach route carries the gorge obstacle described by by Polybius that certifies the southern route identi- Polybius and Livy (Mahaney et al., 2010b). In addi- fied by Sir Gavin de Beer as the only possible en- tion to the gorge and rockfall, all other environmen- trance into cis-alpine Gaul transited by the Punic tal parameters described below line up with argu- Army in the opening phase of the Second Punic War. ments presented previously by Sir Gavin de Beer Identification of the invasion route hinges on the (1956, 1969) and fully support his southern route doublet rockfall (Fig. 2), all other criteria, whether on hypothesis.

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 49

Fig. 2. Traversette rockfall, the main impediment to Hannibal’s exfiltration from the Alps. Figure after Mahaney (2008a). 2. REGIONAL BACKGROUND ly capable of rapid movement, they were unencum- bered by non-combatants and other non-essential The regional background to the environmental personnel, and were used to movement in moun- constraints facing Hannibal and his troops is well tainous terrain. explained in several works, notably, de Beer (1956, Moreover, as previously outlined, the idea that 1969), Proctor (1971), Lancel (1999), Bagnall (1999), the Hannibalic Army moved through a landscape Mahaney (2008a), amongst others. Plunging into the vastly different from today has been disproven by environmental evidence that might be used to target glacial geological (Mahaney, 2008a; Mahaney et al., key sites requiring historical archaeological recon- 2016a) and cosmic airburst (Mahaney and Keiser, struction has been the prime motivation behind 2013; Mahaney et al., 2018b) studies that show the work carried out by Mahaney (2004, 2008a), Ma- landscape seen today is essentially the same as that haney et al. (2007b, 2008c, 2010a, b, 2010b, c, 2014) seen by Hannibal and his troops. The only modifica- involved searching through possible reconstructions tions to the land surface result from minor input of using methods from several disciplines, including aeolian-influxed sediment to soils and paleosols, mi- geology, geomorphology, topography, astronomy, nor stream incision, and increase in the thickness of ecology, philology, palynology, chemistry, and geo- mass wasted deposits emplaced in the form of soli- chemistry. This differs substantially from other earli- fluction deposits, debris flows, and rock and snow er attempts to identify the route, some aimed at avalanches. Col elevations have changed little since naming the route through a local municipality simp- Hannibal‘s time mostly with slight tectonic shifts ly to satisfy tourist activity (Proctor, 1971). Others raising height by a few meters. centered on topography and the elevation of local cols supported by previous military strategists, Na- 3. RESULTS AND DISCUSSION polèon for example, that Hannibal had to have used 3.1 INITIAL RESEARCH STAGE a particular route. Conversely some routes were considered just too difficult for Hannibal to have negoti- Combining expertise as an environmental scientist ated steep inclines and narrow ledges to move his working in various mountain environments on all animals and soldiers into Italia. Many workers seem continents over several decades, coupled with an to forget or overlook that Hannibal had a standing interest in classical history, it is not unexpected that professional army and that his soldiers were not on- Hannibal‘s invasion over one of the most imposing

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 50 W.C. MAHANEY

mountain chains on Earth might pique my interest in agents, Hannibal probably fixed on the lowest of all focusing on just how and where the invasion was passes across the Col de Genèvre (~2000 m), general- carried out. Hannibal‘s feat of moving large numbers ly considered a main Gallic artery, later named of troops, support personnel, horses, mules, let alone Pompey‘s Pass when he led Roman armies into Gaul elephants, across elevations up to 3000 m has few during the Civil Wars. rivals including such figures as Caesar, Belisarius, If one assumes Hannibal intended to take the Du- Charlemagne, Napolèon, even including such mod- rance route to the Col de Genèvre he would have ern figures as Rommel in the Alps and the American had to either swing his army eastward from the 10th Mountain Division in the Aleutians. Only Napo- Rhône to take the Durance north or follow the Rhône lèon comes close to the number count of the Hanni- to the Drôme confluence stopping near Orange to balic Army, and he, it seems never ventured close to meet with lowland Gauls whom Polybius (III, 49) the 3000 m mark. The question of the route itself has affirms he had made a previous arrangement for been the driving force of historians and classicists resupplying his army. Thus, it is highly unlikely that who usually quote each other in constructing the army would resupply at Orange and trek north- ‗proofs‘ of the invasion route, very few ever visiting ward to follow the Isère route and even so, if this the approach routes and cols in question. Up to the route is preferred, one is left wondering where the time of Mahaney (2004, 2008a) no one attempted to first mountain pass might be located before encoun- target key sites where historical archaeological ex- tering the long defile along the final approach route ploration might reveal artifacts related to the inva- to the high col. Further, if, as Polybius (III, 50) as- sion of Hannibal or the later crossing of his brother, serts, this first obstacle had to be taken in a night Hasdrubal, in 207 BC. Motivation to define the route, operation, the Gauls abandoning the high pass; the some consider to be the great question of antiquity only possible cols are in the Dauphiné Alps—Col de (personal communication, T. Corey Brennan, 2004), Cabre (1193 m asl) and Col de Grîmone (1318 m seems to lie in the route designation itself, not to tar- asl)—and only the Grîmone pass fits the topographic get key sites worth historical archaeological explora- assessment offered by Polybius (Mahaney, 2008a). tion. This seems a strange quest to me as given the Once across the Col de Grîmone and into the Du- fact that the Hannibalic Army emerged onto the Po rance valley with local tribes in tow, Hannibal was River plains, the route itself of little consequence ex- faced with mounting obstacles, the followers and the cept that a route designate itself narrows the area of Durance crossing. The followers gave up hostages possible archaeological exploration. Moreover, doc- and cattle but Polybius notes their friendship was umentation of key sites to explore might lead to the uncertain and no doubt his scouts probed both the recovery of important information to elucidate the Guil passage and the Col de Genèvre. We shall never military culture of ancient Carthage. Hence, the en- know why Hannibal diverted from the Genèvre to vironmental targets soon came to carry prime im- the Guil Valley, but if the Allobroges held the for- portance in this forensic exercise. mer, it is certain he would have sustained heavy losses trying to force passage. Polybius is certain that 3.2 ENVIRONMENTAL MATRIX Hannibal was with his rear guard (often his position Reading Polybius and Livy, one comes away with on the trail) making it likely, Maharbal, his second- a certain briefness attained by both historians de- in-command was with the vanguard and may have scribing the invasion, most probably because few made the decision to follow the Guil River. Appar- could write and most place names were spoken only, ently the army deviated from the Durance following maps way off in the future. That Hannibal probably the Guil River into the major defile described by Po- had sketch maps made by his agents who crossed lybius, an 18-km long gorge with steep cliffs where the Alps into northern Italia, it is a given that he the army was trapped by an assault on the rear eche- probably knew in advance the positions of most ma- lons and from the flanks, apparently with the ele- jor routes, including those that sourced from the Du- phant corps and the cavalry up front, Gallic merce- rance and Isère valleys. We may only suppose he naries in the middle and Hannibal with African and had this information but it is certain that being the Spanish contingents in the rear. It was the Gauls commander that he was he would not venture forth who got the worst of the assault (Mahaney and Tri- with an army without knowing his probable route. cart, 2008), which lasted for a day, Hannibal re- Most certainly, the only map of the Mediterranean grouping his troops near a ‗certain bare rock‘ Polyb- (Strabo, trans. Jones, 1999) made two centuries after ius, III, 53) just inside the gorge entrance. The only the invasion is nearly blank north of Massillia (Mar- large rock thus located is within 2 km of the gorge seilles) with a distorted Adriatic oriented nearly entrance as described by Mahaney (2008). However east-west (see Mahaney, 2008a, Fig. 2.2). Clearly be- as described by Polybius (III, 53), with the flanks un- yond probable sketch maps and recollections of protected the Allobroges carried out an enfilade at-

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 51

tack hurling boulders and other loose debris from and scrivener, thin with particulars as they are. The- the heights onto the soldiers transiting the gorge and se parameters include the following: with terrifying effect. Because this account is the on- First, the crossing of the Durance and decision to ly one matching the present topography with the follow the Guil River, if made by Marhabal, follows ancient texts, it is probable that Hannibal marched from assessments by Basil Hart (1967) who mentions from the Drôme to the Durance and into the Guil the unusual authority delegated by Hannibal to basin. Following the initial engagement in the lower Marhabal who often led the vanguard. As previous- foothills, the only such large and long gorge is the ly pointed out, we shall never know who made the Combe de Queyras as described by Mahaney decision to tackle the highest route but it is certain (2008a), all other approach routes leading to Mt. Ce- that Hannibal did not want to fight the Allobroges nis and Col Clapier lacking a long narrow defile (see (Polybius, III, 52). The defile leading to the high Mahaney et al., 2008c, 2010c).Putting the two main slopes, following a lengthy survey of all possible sources of the invasion together, and considering approach routes (Mahaney et al., 2010b) targets only that Livy never left Padua and wrote 180 years after one candidate, the Combe de Queyras (Mahaney, the invasion, while Polybius followed the invasion 2008a), the only gorge with the right length of ~18 67 years after the event, following interviews with km. No other approach route offers the same topo- people who, he says, had taken part in events, Polyb- graphic obstacle with steep walls providing the ius‘ account is preferred. Polybius likely traveled Gauls with a rock source to be hurled at the Cartha- quite possibly with sketches from Silenus in hand, ginians (Mahaney, et al., 2010b). and it is he who provides the most convincing topo- Second, taking snow accumulation into account, graphic, environmental, and possible stratigraphic assuming the crossing was in October, 218 BC, it is information related to the mountain crossing. not unusual in the Alps. The snowline was probably Whereas the topography is described in general similar to today (Neumann, 1992) but a sudden terms by both historians, it is Polybius who provides snowfall may have produced melting followed by more detail on the approach defile, snow, frozen freezing at the ground/snow interface which would ground, view from the col, the rockfall blockage of make footing difficult. Such a snowfall might extend the exfiltration route and the regrouping area below as low as 2000 m which indicates the army‘s first the rockfall. It is Polybius‘ rockfall description of a contact occurred above the long defile, most proba- two-tier rock system that places him within the ranks bly near the present Chateau Queyras. It is possible of the stratigraphic elite of history as he clearly un- higher up that soldiers slid on sporadic permafrost derstood that one deposit was older, the second but this is uncertain as no permafrost was encoun- younger. In translation, at least, he is quoted as in- tered during later investigations (Mahaney et al., terpreting one deposit overlying another, a strati- 2007a; Mahaney, 2008a). Polybius mentions (III, 54) graphic succession clearly seen even today based on that following the assault in the defile the time was weathering differences displayed on cobbles and near the setting of the Pleiades, which meant early boulders in the mass wasted deposits (Mahaney, November. Polybius (Scott Kilvert, trans, 1979) men- 2008a; Mahaney et al., 2014). tions that this could mean even the end of Septem- Polybius was presumably an accomplished mili- ber, 218 BC, but snow could start this early in fall. As tary officer but his observation places him within a pointed out by several authors, including Polybius, group of extremely keen interpreters of the land- the crossing is documented to have occurred at the scape. Few, even today, would venture across the setting of the Pleiades (Walbank, 1956; Polybius, Traversette rockfall and recognize two superposed 1979; Livy, 1965), the astronomical record allowing rock deposits distinguished principally on weather- that the Carthaginians were favored by a ―Hunter‘s ing features and soil morphogenesis. Perhaps such Moon‖ (Full Moon on October 30; Proleptic Julian weathering features as color and oxidation states of Calendar, documented by Prof. D. H. Seaquist, Uni- clasts in the two deposits were likely more apparent versity. of Toronto in Mahaney et al. (2008c) which in Polybius‘ time than today since the youngest mass would have provided sufficient moonlight for the would likely have lacked major lichen cover and ox- descent on the Italian side to have taken place on a idation effects would have had an age of only a few twenty-four hour regimen. The moon‘s elevation centuries if the observations of Mahaney (2008a) are above the horizon at the time of transit (about 54 correct. degrees at this site) was the highest for the year, fa- Thus, key environmental parameters comprise an vorable for illuminatiion of the rugged surface and environmental matrix, construed from Polybius and deepening valley of the descent into the upper Po. Livy, the former relying on those works available to Nevertheless, single file evacuation of ~20,000 infan- him from afar including most likely the journals of try, ~7,000 cavalry (horses plus men), ~10,000 draft Silenus who accompanied Hannibal as topographer horses/mules, ~5,000 handlers and 37 elephants

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 52 W.C. MAHANEY

would require an estimated ~200 hr (i.e. 6-8 24-hour river plains as the exfiltration point, thus excluding days) to clear the col and upper valley, which is the Dora Riparia leading to the Po from the Clapier within the estimated three days to clear a path and Cenis cols some 50 km to the north. Polybius through the rockfall (Polybius, III, 55). Whereas we (III, 54) is said to have called his men to the col and cannot be sure of the troop/animal numbers outside pointed the way toward Rome to reassure them, but of the elephants, we do know that Hannibal he could not have amassed the entire army on any of emerged from the mountains with 26,000 soldiers, a the cols, and the Col de la Traversette itself could figure documented by Polybius (III, 33) from the barely hold 100 soldiers, so he must have addressed bronze tablet in the Temple of Hera at Cape Lacini- his key lieutenants not the entire army. um, a summary considered correct by most modern Fifth, the rockfall in question can only be found workers (de Beer, 1969; Lazenby, 1998; Lancel, 1999). below the Traversette col, a mass of rubble matching Thus, with this number of Carthaginian infantry and closely to Polybius‘ estimate of 250 m wide and cavalry and 37 elephants, it is reasonable to assume comprised of two stages of development, one older his baggage train contained at least several thousand and the other and upper one younger. While Livy horses and mules for which an unknown number of mentions that Hannibal fired the rockfall to heat and handlers would be needed. Adding in cooks, artifi- split boulders, Polybius is mute on the subject, indi- cers, armourers and engineers, the entire army prob- cating the firing event never happened. Further, fol- ably numbered (considering losses during the as- lowing an extensive investigation of the rockfall sur- cent) a number in excess of 26,000 soldiers emerging face, clasts embedded in the deposit surface did not onto the Po River plains. reveal any vestige of recent firing, no carbonized Third, regrouping area in the upper valley, uni- material (Mahaney et al., 2008c, 2010b). It appears dentified until Mahaney et al., 2017a, 2017b, 2017c, the rockfall firing episode described by Livy is mere discovered churned-up beds in alluvial mire sedi- fiction, and more to the point, where and how would ment (Fig. 3) at 2800 m dated to within the Hannibal wood be obtained as firing material by soldiers close window of 2168 cal yr BP. These beds (Mahaney et to starvation (Bagnall, 1999) and with timberline far al., 2017a), at approximately ~40 cm depth, with ~15 below the rock mass. cm amplitude, contain faecal biomarkers [(5β- Sixth, the regrouping area could lie below any one stigmastanol and the bile acid deoxycholic acid of a number of possible cols of passage but the likely (DCA)] (Mahaney et al., 2017b) related to horse and area below the Traversette is on a bedrock plateau mule manure with roundworm eggs and bacteria, floored with mire and coalescing alluvial fans (Fig. the latter belonging to microbe groups of Firmacut- 4) at ~2000 m asl, the latter of which reveal a ies and Clostridia related to gut release from horses churned-up bed similar to the alluvial mire on the and mules (Mahaney et al., 2017b). These findings French side (Mahaney et al., 2018a). That the lee side show the passage of thousands perhaps tens of thou- area has been used for centuries is obvious from a sands of animals and potentially humans. Studies of stone house, complete with animal quarters and a the magnetic susceptibility of sediment within the urine trough oriented north south, draining onto French mire showed a slight increase in the churned- slopes on the north flank of the Po river and stone up beds of several sections (Mahaney et al., 2017c) walls at around 2250 m used to pen animal stock as which may relate either to abnormal concentrations described by (Mahaney et al., 2010a). of magnetite or to the presence of weathered arti- Seventh, most historic arguments have mentioned facts. In any case the mire should be subjected to a Hannibal‘s famous elephants but few have ques- metal detection survey and possibly to GPR (ground tioned the ability of elephants to make the journey penetrating radar) investigations. As well, lowland over the Alps. John Hoyte (Hoyte, 1960) of Oxford areas behind the Younger Dryas moraines adjacent University experimented with an elephant crossing to the Guil mire should also be subjected to similar Mt. Cenis showing that it was possible and that surveys. Hannibal could easily have managed at least the Fourth, the col from which Hannibal looked crossing to the north at lower elevation. However, across the plains of the Po is one of the main envi- elephants are known to work at high elevations as ronmental parameters cited by historians. Investiga- documented by Mahaney (1990) reaching elevations tions at all possible crossing places indicate the above 4500 m. Thus, crossing the Traversette col at Traversette col is the only passage which gives a ~3000 m asl would have been possible although dif- long view, on a clear day nearly to Milan (de Beer, ficult given the narrowness of track in place and lack 1969). The two cols adjacent to the Traversette—Col of forage. The remains of Hannibal‘s elephants have Agnel and Col de la Croix—offer restricted views been found in the Arno terraces near Florence (Ro- cross country and do not look down upon the Po mano and Palombo, 2017). plains. Because Polybius definitely targets the Po

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 53

Fig. 3. French mire at 2850 m astride a tributary to the Guil River (see Mahaney et al., 2017c). Areas like this, although scarce, are considered to provide water and forage to Hannibal’s projected route into Italia (cis-alpine Gaul). The Younger Dryas moraine and surrounding glacial landscape set in place during the last ice age has been little modified right up to today. Hence, despite claims to the contrary, the landscape viewed by the Punic Army is the same as exists today. Photography by Pierre Tricart. Eighth, dating the evidence required using relative ada Soil Survey Committee (1998) and the USDA vs. absolute methods. Relative age-dating methods rely (Soil Survey Staff, 1999). All soil and sediment colors on criteria that relate one topographic/geomorphic reported in the literature are based on the color chips entity against another using, for example, height of of Oyama and Takehara (1970). one terrace over another (ex. Mahaney et al., 2016b), The Traversette rockfall, a case in point, illustrates the difference in elevation a measure of the erosion how surface weathering features (i.e. color of surface event followed by a deposition (filling process). In clasts) resulting from physical and chemical altera- other cases, changes in strength of weathering and tion that distinguishes the two deposits could be soil development have been used to establish rela- used to separate deposits in time (Mahaney et al., tive ages in many alpine areas (ex. Mahaney and 2014). In this case we know the older lobe of debris is Kalm, 2013). Also, if different in time the older de- >12.8 ka because it contains cosmic material correlat- posit is expected to carry stronger pedogenesis, that ed to the black mat event (Mahaney and Keiser, is, a more complex weathering profile compared 2013) and the paleosol within can be correlated to with the younger deposit. One cannot say for certain Late Glacial (15 ka-12.8 ka) sediment); the younger what time is involved, only that a soil (e.g. paleosol) lobe contains less weathered clastic materials related in the older terrace represents greater time for de- to weathering and soil development considered to be velopment. Also, hearths in deposits that may have of middle Neoglacial (~3.0 ka) age; thus the sedi- been used as bivouacs for a long time may contain ment seen by Polybius must be >2.2 ka. Polybius, of stratigraphic evidence belonging to ecologic disturb- course was most likely guided by the color differen- ance, possibly the Hannibal event, and may contain tial which clearly delineated the two deposits on the charcoal material that could be dated by AMS C14. basis of chemical alteration products (i.e., oxides and Where reported, all soil descriptions follow the Can- hydroxides) and illustrates his extraordinary powers

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 54 W.C. MAHANEY

of observation. My interpretation of the rockfall and there is little supporting evidence (Mahaney, 2008a). its importance in the route interpretation was criti- The evidence for the Traversette rockfall as the key cized by Kuhle and Kuhle (2012) who favored along site supporting the southern route of de Beer (1969) with others the northern route via the Col de was summarized by Mahaney (2013). Clapier, an entrance to cis-alpine Gaul for which

Fig. 4. Upper Po valley, with Pian de Re albergo off to the left, and mire and coalescing alluvial fans showing core and section sites V21, V21A and V21B (in center). Photography and site annotation by Peeter Somelar; figure after Mahaney et al. (2018a). Core and section data from these sites is interpreted in Mahaney et al. (2018a). To obtain absolute ages on materials in this case re- dates, with somewhat smaller standard deviations lies only on radiocarbon (Mahaney et al., 2017a, b, for the latter method. All AMS radiocarbon analysis and c), which is used to date relict organic material and date assignment follow Bronk Ramsey (2009), buried in section. Because the half-life of the 14C iso- Bronk Ramsey and Lee (2013) and Reimer et al. tope is 5710 yrs most of the recovered organic mate- (2013). rial relevant to solving the Hannibal enigma is well An inventory of possible sites for radiocarbon da- within the first one-half of the first half life decay ting along the targeted routes produces alluvial ter- process, which means that providing there is no con- races covered with surface soils and buried pale- tamination of samples one can expect accurate dates osols, alluvial fans along valley sides, alluvial mires, with low standard deviations. However, since rele- and possible hearths. In most alpine areas of the vant samples are within ~50 cm depth there is al- world bogs and lacustrine plains often provide a ways the possibility of leaching of younger organic means of obtaining bog-bottom 14C dates which are, carbon downward in section, thus producing at best, minimum ages on organic remains in depos- younger ages and wider standard deviations. Cou- its within which such materials are found. Once a ple to this, the possibility of ground water influx into depression is formed it may infill with inorganic sed- older sediment, the addition of younger organic car- iment for a time (ecesis period), before which recog- bon as disolved organic matter may produce the nizable organic remains accumulate which can be same effect. We may get around this by careful scru- dated; hence, the minimum age designated. Lacus- tiny of the standard deviations for each date to in- trine plains, reacting to lake level fluctuations may sure the standard deviation is <3% or approximately contain buried soils that may be dated by 14C but ~60 yr or less. The maximum age possible for con- these are not present in suitable positions to add to ventional radiocarbon dates is ~40 kyr; this increases the Hannibal story. To add hearths to the list re- to ~70 kyr for AMS (Accelerator Mass Spectrometry) quires one to overcome the speculation that requires

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 55

one to think that Hannibal‘s army would not pass up 1/1, 2/1; 7.5YR 1/1, 2/2, 2,2; to 5YR 1/1,2/1. YR a site like this and some assurance that Hannibal refers to hue, the dominant wavelength, where de- actually took the route specified. The same applies to creasing color refers to strengthening of the yellow alluvial terraces and with these surfaces the prospect red color. The two numbers that follow refer to bril- of organic retention is not great which means large liance with increasing value from dark to light, and samples are required to obtain even minimum or- chroma, purity of dominant wavelength, which in- ganic residue that might be obtained for dating. Al- creases with decreasing white light. Common wave- luvial mires, both in the upper Guil and Po catch- lengths for young sediment such as in the upper Po ments, present the best options for dating because Valley are 10YR (yellow red) and 2.5Y (yellow), the water velocities are low and retention of weathered former for various degrees of weathering and the organic sediment is great combining to preserve dis- latter for fresh, unweathered sediment. To use the turbed sediment beds from some previous episode analogy of a cube to which hue, value and chroma of bioturbation such as disturbance from a foraging are measured along three edges (Millar et al., 1966), army. In the case of the upper Po drainage the pres- each color notation represents a point in the cube, ence of an endangered species of salamander re- i.e., the Munsell notation. Within each major hue quired that all sampling be relegated to alluvial fans represented here, a lower value and chroma repre- away from mire sediments that might have yielded sent a darker and more carbon-rich sediment. For better results both in terms of the physical and strat- lithology in the field area, I relied on geological igraphic evidence (Mahaney et al., 2018a). mapping by Tricart et al. (2003). Pulling cores and digging sections in alluvial sed- One of the great assets of working in the Guil and iment produces planar beds aligned horizontally Po valleys rests on the detailed bedrock mapping with the earth‘s surface so that variance in flow re- (Tricart et al., 2003) carried out by Pierre Tricart and gime and or changes in source sediment produce his associates from the Institute of Geology, Univer- sorted sediment even in beds with high organic in- sity of Grenoble. Having bedrock control for all sites put. Cores of 50-70 mm diameter limit the amount of explored in this area insured that lithologic relation- recovered material for dating and chemical analyses ships could be worked out in the laboratory prior to whereas sections produce open faces with widths of engaging in field work. 50 cm + allowing collection of material for replicate Finding the churned-up bed is one part of the dating where necessary and higher weight material task, dating the transition zones to younger beds up for laboratory analyses of various kinds. The analyst section and older planar beds down section, and ex- expects to collect downward from the surface into plaining the greater than 2168 cal yr BP dates within progressively older beds (shown below in Fig. 5) and the bioturbated mass still another. The boundaries of if bioturbated (churned-up) beds are encountered the bioturbated mass are so diffuse that the relevant such disturbed beds will produce wavelengths and radiocarbon dates may not truly reflect start up and amplitudes of distorted beds with various dimen- shut down of the bioturbation process even though sions, physical evidence normally not seen in cores. the two dates should be approximately the same, In the case of the French vs. the Italian sediment that is, representative of the x number of days re- complexes, one might expect the softer French sedi- quired to produce the bioturbated bed. To further ment to deform to a greater degree than the Italian complicate the dating process it is necessary to pro- coarser sediment, which is approximately what hap- duce a string of dates from near the surface to ap- pened. The French sediment sites showed an approximately the top of the churned-up bed which is proximate center of disturbance at 40 cm ± 15 cm a hit or miss proposition as shown in Fig. 5 that re- amplitude whereas the Italian sites showed disturb- lates to section V21A in Mahaney et al. (2018a) ance centered at 45 cm ± 10-15 cm in two out of three where the upper boundary is ~40 cm, the lower sites; 60 cm ± 15 cm in the third site. The variation in boundary at 55 cm but with some variance given the depth of disturbance is explained by differences in magnitude of disturbance. sedimentation rates partly the result of a wetter cli- 3.3 INTERPRETING THE PROOFS mate in Italy compared with France. Aside from the loss of planar beds nearly horizon- The lead up to the only possible defile along the tal to the Earth‘s surface that is common to mire and approach route, the Combe de Queyras, is along ei- alluvial fan deposits (Mahaney, 1990), sediment ther the Drôme River from the Rhône or from the within the bioturbated beds is commonly much south along the Durance. The southern route of darker (addition of faecal matter and reworked hu- some 150 km from the Durance/Rhône confluence mus) than variegated colors encountered up or does not lead to a col described by both Polybius down section from the disturbed beds. Common and Livy but the Drôme approach does end at the color ranges within the bioturbated beds are: 10YR Col de Grîmone in the Dauphiné Alps, the topogra-

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 56 W.C. MAHANEY

phy of which matches the landscape where Hannibal Gauls unleashed a ferocious attack that nearly over- himself led a night operation to secure passage. The- ran the rear of the army (Polybius, III, 53). Because se lowland routes have been thoroughly discussed the elephants and cavalry were up front it was the by de Beer (1969) and Proctor 1971), as to likely di- Gallic mercenaries who caught the brunt of missiles versions, but the Drôme route would take Hannibal and rocks hurled from the high ridges astride the close to the Col de Genèvre (~2000 m), his likely in- Combe, the result being that Hannibal suffered tended col of passage. heavy losses from what appears to be an enfilade As Hannibal‘s army approached and crossed the attack a tactical maneuver similar to what he himself Durance, Hannibal was followed by the Allobroges later used at Trasimene against the Romans in the who threatened his rear echelons (Polybius, III, 52). Apennines (Mahaney and Tricart, 2008), a more The question of why the Carthaginians diverted modern example coming from Lee‘s attack on the through the Guil valley from the Durance instead of Union Army at Gettysburg during the American marching towards the Col de Genèvre is uncertain Civil War. but what is certain is that Hannibal was with the rear Investigations carried out in the Combe de Quey- guard and the army, often led by Maharbal (Hart, ras over the period 2002-2010 by Mahaney and Tri- 1967) who, according to Polybius, was informed of cart (2008) pinpointed unusually large accumula- other Allobroges massing at the target col. Polybius tions of boulders and cobbles along and at the base is on record saying he did not want to fight the of the northwestern wall with the Guil River to the Allobroges, so it is possible the order to move southeast. It would appear that even in Hannibal‘s through the Combe de Queyras came from Ma- time the approach was to the west of the river but harbal, Hannibal having his energies directed to the recent flood control efforts may have rearranged rear. If this is correct, it is odd, if not unusual, that some of the rubble deposits, so finding artifacts in the flanks were not protected adequately because this mass would prove difficult and probably impos- once into the defile presented by the Combe, the sible.

Fig. 5. Radiocarbon dated beds in section V21A showing the age progression from near the surface to approximately ~40 cm depth, the transition from time of normal sedimentation to the onset of massive ecological disturbance, i.e., the churning event. The >2168 cal yr BP dates represent sediment transported into the churned up morass by animals, humans or natural processes such as surface runoff or aeolian influx. Figure from Mahaney et al. (2018a). Color designa- tions for dark to light beds are as explained above.

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 57

Following the Guil River toward the high col, the 3.4 ALTERNATIVE HYPOTHESES only camping or regrouping place is between 2400 Acceptance of the premise of both Polybius and and 2650 m alongside the main drainage and a tribu- Livy that Hannibal‘s army foraged in the catchment tary that passes the G5 alluvial mire sites (Fig. 3) in- below the two-tier rockfall for a period of five days, vestigated by Mahaney et al. (2017a, 2017b and such activity should have ravaged the upper valley 2017c). To the south of the G5 sites several low-lying and be recorded in the alluvial fan sediment record- flat lacustrine plains offer some forage opportunities ed there. Any existing soil would have been ground that were likely used by the Punic forces as they re- up and mixed with sediment in situ and imported by formed following the battle below in the Combe de animal action. Because the surface soils at all three Queyras. Attempts to section two of these sites to the sites are similar (about 12-cm thickness with A (Ah), south of G5 (Mahaney et al., 2017c) produced incon- C, Cu horizons, classed as Entisols (Soil Survey Staff, clusive results indicating the ecologically disturbed 1999) or Regosols (CSSC, 1998), correlation with sim- beds are likely deeper than ~1 m. If, as Polybius (III, ilar pedons in the upper Guil Valley (Mahaney et al. 54) indicates, the army camped in snow, some of 2016a) suggests these profiles are young and have these depressions were probably used as water been in situ for only two centuries or less. Both soil sources and even today the main Guil river drains suites date to the end of the Little Ice Age (less than the south side of the valley and would have provid- 170 yr), and their existence further suggests the soils ed a valuable source of water. More to the point, have been protected from flooding that may have given the faecal matter and microbe analysis carried occurred since that time. Deeper into the three sec- out at the G5 site (Mahaney et al., 2018b), which in- tions of the upper Po river valley, and the multiple dicates widespread accumulation of animal and pre- sections of G5 in the upper Guil valley of France, sumably human defecation, the entire moraine com- darker beds encountered suggest periodic or aperi- plex at 2400 to 2600 m asl in the upper Guil Valley is odic attempts to produce Histosols or soils with probably laden with metagenomic DNA and fossil dominant organic compounds relative to inorganic faecal matter related to Hannibal‘s Army. sediment. However, because these organic beds lack Reaching the Col de la Traversette at ~3000 m on a oxidized horizons, it appears such attempts, alt- clear day, one can see across the Po plains below to hough possibly long-lived in some cases, did not some distance, possibly as far as Milan, as de Beer produce pedons. (1969) has stated. The usual view from below the Col Alternative testable hypotheses to bioturbation de la Traversette is obscured by clouds as illustrated (churning) of sediment caused by Hannibal‘s Army on the cover of The Warmaker (Mahaney, 2008b). might be: (1) upper flow regime flooding; (2) earth- Other route passages to the north at Col de Clapier quake vibration; (3) transhumance; and (4) soil cry- and Col de Mt. Cenis have limited views from the oturbation. The upper Po Valley is susceptible to summit and all look down upon the Dora Riparia extreme climatic events originating out of the Adri- which joins the Po River where Hannibal sortied atic, as illustrated in 2008 when excessive precipita- against the Roman Army of Scipio whence he made tion induced intensive mass wasting and flooding in contact, first at the Ticinus and latter at the Trebbia the form of debris flows that inundated valleys and before retiring into winter quarters to the south. buried deposits up and down the Western Alps from The main proof for the Traversette Route is the Mt Cenis all the way to the south of Mt Viso (Ma- blocking rockfall at 2600 m, the doublet deposits of haney et al. 2010a, 2010b). Such flooding events pro- such, the oldest lobes dating to at least 12.8 ka be- duced well-sorted and stratified sediment succes- cause they carry the black mat signature (12.8 ka) sions lacking the mixed-sediment character of the seen with microscopic analysis carried out by Ma- bioturbated beds seen in the French mire and in the haney and Keiser (2013) dating the early accumula- upper Po sections described here. Whereas turbulent tion of rubble to the Late Glacial and well before flooding events are possible and cannot be totally Hannibal‘s time. The younger rubble accumulation discounted, the available evidence indicates that bio- that covers nearly three-quarters of the rockfall area turbation is the process that produced the observed is most likely of middle Neoglacial age judging by sediment disruption. As reported by Mahaney et al. weathering features and soil morphogenesis that (2018a), particle size tests on a random collection (n compares favorably with soil research carried out = 6) of small samples taken from the churned-up bed previously by Mahaney (1991). There is no similar in V21A in the upper Po valley, show extreme varia- rockfall deposit on the lee side of the Alps that tions in texture ranging from coarse sands (some- would stop an army, only minor accumulations of times with pebbles) to fine silts and clay with no rock rubble below all other passes (Sodhi et al., sorting visible. Water-sorted sediment recovered 2006). from alluvial/colluvial fans and lake beds (see Ma-

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 58 W.C. MAHANEY

haney, 1990) follow the principle of horizontality, lamina stratigraphy, which is clearly lacking in the that is, beds are parallel with the Earth‘s surface. Un- disrupted beds of the sediment stack reported by less disturbed by high-velocity currents, such as ex- Mahaney et al. (2018a). Similar experiments using treme flooding that occurs with jökulhlaups etc. or vibration tables have illustrated the effects of earth- moraine bursts, such bedding is nearly horizontal quake activity on sediment sorting (Miyamoto et al., with the Earth‘s surface. Exhaustive studies of Holo- 2007), forcing dislodged sediment into discrete cene alluvial fans on Mt Kenya (Mahaney, 1990) groups. Whereas earthquake activity in the Western failed to produce distorted bedding despite the pres- Alps is known to be an ongoing process, the deposits ence of active glaciers releasing meltwater during under investigation here show little sign of full-scale postglacial time. vibration or neotectonic activity. Freeze–thaw processes in resident soils might Competing routes for the Hannibalic invasion are produce convoluted beds or horizons which are not shown in Figure 1 (Mahaney et al., 2017a), although parallel in extracted cores or in dug sections excavat- various other authors (i.e., Guillaume, 1967; Proctor, ed in the alluvial fans. Such convolutions usually 1971) have argued for other cols adjacent to the leave beds or laminae intact, that is, upturned but Traversette and Mt Cenis, without linking topo- lacking a disrupted or mixed quality such as is evi- graphic and geomorphologic evidence to support dent in the churned-up beds described here. In any their claims. Time/motion analyses (Mahaney case, heavy snowfall in the Western Alps provides 2008a) of the northern (Cenis) versus the southern an insulating medium that may counteract freeze– (Traversette) show the Traversette route to be slight- thaw processes in underlying soils (for examples of ly shorter but more rugged over the Traversette Cryosols, see Mahaney and Kalm, 2013). pass, longer over the less rugged Mt Cenis. This as- Transhumance might be expected to produce sed- sumes that Polybius clearly differentiated between iment disturbance and it is an activity known to the Po plains and the plains adjacent to the Dora Ri- have occurred in the upper Po Valley, as in several paria (a northern tributary of the Po River). The only Italian valleys, over several centuries, but no record route leading to the Po plains is over the Traversette. of churned-up or bioturbated sediment exists above All variations between elevations, snowline, pres- the proposed Hannibal beds identified at c.2200 cal ence of frozen ground, long defile on the approach to yr BP in the subject area. Indeed, while transhu- the mountains and a large rock just inside the defile mance has been prohibited in the upper Po Valley entrance have been discussed by Mahaney (2008a). since the Second World War, the surface soil profiles 4. CONCLUSIONS in resident alluvial fans—V21, V21A and V21B—lack any sign of physical disturbance, which supports The various stages of reconstructing the Hannibal similar evidence reported in the upper Guil Valley of enigma illustrate just how information buried in the France, all this despite the ongoing practice of trans- classical literature might be reconstructed to provide humance occurring every year (Mahaney et al., 2016, an environmental matrix that can be tested using 2017a) in the Guil Valley. Linking the churned-up various scientific analyses. In this case, work started beds in cores and sections in the upper Po Valley with a topographic and geomorphologic base that with the Hannibal event of 2168 cal BP means there could be used to provide time/motion analysis to is a time correlation between the churning event and establish movement of the Punic Army, results indi- the reconnaissance of the Hannibalic route in the cating that of the three proposed routes outlined in upper Po Valley, Italy‘s postulated passage of Han- Fig. 1, only the middle route is the longest, the nibal. However unlikely, the coeval relation between northern and southern routes almost identical in the churning event and high-resolution dating may time distance but not regarding physical obstacles. relate to some other anthropomorphic event not rec- The southern route across the highest of all passes orded in history. The churning event may also be with the most rugged terrain would require extra partly related to menacing Gallic tribes following time but just how much is unknown. The battle in Hannibal after the altercation in the 18-km defile, as the approach defile and construction through the described by Polybius (Polybius/Scott-Kilvert, 1979, rockfall placed heavy burdens on the army, especial- III, 53; surface analysis by Mahaney, 2008a). It rely since as Polybius indicates they lost much of their mains for DNA analysis to link firmly recovered mi- provisions and this must have put a strain on ani- crobes and helminth parasite eggs with Spanish mals, principally upon the elephant corps. and/or North African animals. Beyond the initial phase, expertise in a variety of While extremely sidelined as a probable cause of fields was required to put numbers to the allotted the churned-up beds, vibrations caused by earth- time of crossing requiring mathematical/statistical quake activity might be expected to produce dislo- studies and astronomical calculations as to the cated materials but with bed-to-bed or lamina-to- amount of moonlight, and timing of movement with

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 59

the hunter‘s moon. Excavations of mires and hearths the Traversette has been known as ‗Hannibal‘s Pass‘ required pedological, geochemical, radiocarbon, and even during WWII, SOE (Special Operations palynological and microbiological inputs from a va- Executive of British Intelligence) referred to Mt. Ce- riety of co-authors, all of which are listed on numer- nis as Napoleon‘s Route, the Traversette as Hanni- ous papers in reference lists as noted in this paper. bal‘s Route in secret documents recently made pub- While permits were obtained to carry out the geolog- lic (Mulley, 2012). ical and biostratigraphic work, no attempt was made Polybius is also on record using the phrase—tas to obtain archaeological input beyond what is out- hyperbolas tas anōtatō tōn Alpeōn—the highest pass of lined in various papers. The only artifacts obtained the Alps (personal communication, J. Lazenby, 2007). are in the form of microbes and faecal matter ob- Thus, whatever your historic theorem of possible or tained from sections dug in sediment piles in various impossible routes, forensic analysis of all environ- places and from cores taken in similar settings. The mental parameters in the ancient literature is heavily main findings of this research affirm the interpreta- weighted in favor of Sir Gavin de Beer‘s southern tions of Sir Gavin de Beer some decades ago that route across the Traversette. Despite the single file Hannibal for one reason or another took the highest approach off the col to the permanent snowbank and most difficult route into Italia across the Col de above the discarded Italian army barracks at ~2650 la Traversette. Indeed, even Polybius is on record m asl, the only route assigned by key environmental saying that Hannibal took the highest route, later parameters is the Traversette negotiated with diffi- named in order of position south of the Col de Ge- culty by the Hannibalic Army. nèvre, Hannibal‘s pass. For the last two millennia, ACKNOWLEDGEMENTS I thank the National Geographic Society (Grant no. 9988-16) for funds to carry out the field and dating part of this research and Lion TV (London) and ZDF (Germany) for filming and allocating funds to Quaternary Surveys during July, 2017. P. Somelar, member of the 2015 and 2017 field teams, was funded also by Estonian Research Council grant PUT1511P. A number of colleagues contributed to this work over the last 15 years and are too numerous to name in full. In particular, I am indebted to Pierre Tricart (Grenoble) for his invaluable assistance in the field and for providing liaison with the Guil Valley Regional Park authority. Chris Allen (Queens University Belfast) and his team provided invaluable microbiological evidence. Brian Kelleher (Dublin City University), and his associates, aided with chemical analyses of samples, and Allen West (Comet Research Group) assisted with advice on Ox Cal radiocarbon dating. Randy Dirszowsky assisted in the field in 2011 and 2017. Marco Rastelli assisted with research permits to carry out field work in the Mon Viso National Park, Italia. REFERENCES Bagnall, N. (1999) The Punic wars, London, Pimlico. Bronk Ramsey, C. (2009) Bayesian analysis of radiocarbon dates, Radiocarbon, Vol. 51, pp. 337–60. Bronk Ramsey, C., and Lee, S. (2013) Recent and planned developments of the program OxCal., Radiocarbon, Vol. 55(2–3), pp. 720–30. Brown, J. E. T. (1963) Hannibal’s route across the Alps, Greece and Rome, Vol.10, pp. 38–46. Canada Soil Survey Committee, 1998. The Canadian System of Soil Classification, 637, Ottawa, Canada, NRC Research Press, pp. 187 (Publ. 1646). Connolly, P. (1981), Greece and Rome at war, London, Greenhill Books. de Beer, Sir G. (1956) Alps and elephants, N.Y., E. P. Dutton. de Beer, Sir G. (1967) Hannibal’s march, London, Sidgwick and Jackson. de Beer, Sir G. (1969) Hannibal: challenging Rome’s supremacy, N.Y., The Viking Press. de Beer, Sir G. (1974) Hannibal: the struggle for power in the Mediterranean, London, Book Club Associates. De Montholon, J.F.T. (1905) Momoires a Sainte Hélène, v. IV, Paris, Garnier, pp. 277-281. Dodge, T. A. (1891) Hannibal, Boston, Houghton-Mifflin., Freshfield, D. W. (1886) The Alpine pass of Hannibal, Proceedings of the Royal Geographical Society and Month- ly, Record of Geography, Vol. 8 (10), pp. 638–44. Freshfield, D. W. (1899) Hannibal‘s pass, The Geographical Journal, Vol. 13 (5), pp. 547–51. Guillaume, A. (1967) Annibal Franchit les Alpes, 218 J.-C., Cahiers de L‘Alpe, La Tronche-Montfleury. Hart, B. H. L. (1967) Strategy, London, Faber and Faber. Hoyte, John (1960). Alpine Elephant in Hannibal‘s Tracks, Eugene, Oregon, Wiff and Stock, 190 pp.

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 60 W.C. MAHANEY

Kuhle, M., and Kuhle, S. (2012) Hannibal gone astray? A critical comment on W. C. Mahaney et al.: ‗The Traversette (Italia) rockfall: geomorphological indicator of the Hannibalic invasion route‘ (Archae- ometry, 52, 1 [2010] 156–72), Archaeometry Vol. 54, pp. 591–601. Lancel, S. (1999) Hannibal, Oxford, Blackwell. Lazenby, J. F. (1998) Hannibal’s war, Norman, University of Oklahoma Press. Livy (1965) The war with Hannibal, transl. A. de Sélincourt, London, Penguin. Mahaney, W. C. (1990) Ice on the Equator, Ellison Bay, Wisconsin, Wm Caxton Press. Mahaney, W.C. (1991) Later Pleistocene and Holocene glacial chronology at Chamonix and Argentière, French Alps, Zeitschrift für Geomorphologie, Vol. 35 (2), pp. 225-237. Mahaney, W. C.( 2004) Geological/topographical reconnaissance of Hannibal‘s invasion route into Italia in 218 BC, in Studies in military geography and geology (eds. D. R. Caldwell et al.), pp. 67–78, Kluwer Academic, Dordrecht. Mahaney, W. C. (2008a) Hannibal’s odyssey: environmental background to the Alpine invasion of Italia, Pisca- taway, NJ, Gorgias Press, , pp. 221. Mahaney, W.C. (2008b) The Warmaker., Bloomington, Indiana, iUniverse, , pp. 302. Mahaney, W. C. (2013) Comments on M. Kuhle and S. Kuhle (2012): ‗Hannibal gone astray? A critical comment on W. C. Mahaney et al., ―The Traversette (Italia) rockfall: geomorphological indicator of the Hannibalic invasion route‖ (Archaeometry, 52, 1, [2010], 156–72}‘, Archaeometry, Vol. 55, pp. 1196– 1204. Mahaney, W.C. and Kalm, V. (2013) Late Holocene paleoclimate and weathering history in the Norra Storfjället Mountains, Sweden, deglacial record and soil stratigraphy, Geografiska Annaler, Vol. 95, pp. 145-158. Mahaney, W. C. and Keiser, L. (2013) Weathering rinds—unlikely host clasts for an impact-induced event, Geomorphology, Vol.184, pp. 74–83. Mahaney, W.C. and Tricart, P. (2008) The unknown Gallic commander: Hannibal‘s debacle in the Combe de Queyras in 218 BC. Nathanail, C.P., Abrahart, R.G., and Bradshaw, R.P., eds. Military Geography and Geology: History and Technology. U.K., Land Quality Press. Mahaney, W. C., Miyamoto, H., Dohm, J., Baker, V., Cabrol, N., Grin, E., and Berman, D. (2007a) Rock glaciers on Mars: Earth-based clues to Mars‘ recent paleoclimatic history, Journal of Planetary and Space Sciences, Vol. 55 (1–2), pp. 181–92. Mahaney, W. C., Milner, M. W., Sodhi, R. N. S., Dorn, R. I., Boccia, S., Beukens, R. P., Tricart, P., Schwartz, S., Chamorro-Perez, E., Barendregt, R. W., Kalm, V., and Dirszowsky, R. W. (2007b) Analysis of burnt schist outcrops in the Alps: relation to historical archaeology and Hannibal‘s crossing in 218 BC, Geoarchaeology, Vol. 22 (7), pp. 799–818. Mahaney, W. C., Kalm, V., and Dirszowsky, R. (2008a) The Hannibalic invasion of Italia in 218 BC: geological/topographic analysis of the invasion routes, in Military geography and geology: history and technology (eds. C. P. Nathanail et al.), pp. 76–86, Nottingham, Land Quality Press. Mahaney, W. C., Kapran, B., and Tricart, P. (2008b) Hannibal and the Alps: unraveling the invasion route from the geological evidence, Geology Today, Vol. 4 (6), pp.223–30. Mahaney, W. C., Kalm, V., Dirszowsky, R. W., Milner, M. W., Sodhi, R. N. S., Beukens, R., Dorn, R. I., Dorn, R. I., Tricart, P., Schwartz, S., Chamorro-Perez, E., Boccia, S., Barendregt, R. W., Krinsley, D. H., Seaquist, E., Merrick, D., and Kapran, B. (2008c) Hannibal‘s trek across the Alps: identification of sites of geoarchaeological interest, Mediterranean Archaeology and Archaeometry, Vol. 8 (2), pp. 39–54. Mahaney, W. C., Barendregt, R. W., Carcaillet, C., Tricart, P., Rabufetti, D., and Kalm, V. (2010a) Debris flow burial of ancient wall systems in the upper Po River valley, Italia, Geology Today, Vol. 26 (6), pp. 209–15. Mahaney, W. C., Tricart, P., Carcaillet, C., Blarquez, O., Ali, A. A., Argant, J., Barendregt, R. W., and Kalm, V. (2010b) The Traversette rockfall: geomorphological reconstruction and importance in interpreting classical history, Archaeometry, Vol. 52, pp. 156–72. Mahaney, W. C., Kapran, B., Kalm, V., Tricart, P., Blarquez, O., Milner, M.W., Barendregt, R.W., and Some- lar, P. (2010c) Hannibal‘s invasion route: age-old question revisited with new scientific data, Ar- chaeometry, Vol. 52, pp. 1096–109. Mahaney, W. C., Keiser, L., Krinsley, D. H., Pentlavalli, P., Allen, C. C. R., Somelar, P., Schwartz, S., Dohm, J. M., Dirzowsky, R., West, A., Julig, P., and Costa, P. J. M. (2013) Weathering rinds as mirror images of palaeosols: examples from the Western Alps with correlation to Antarctica and Mars, Journal of the Geological Society, Vol. 170, pp. 833–47.

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 THE HANNIBAL ROUTE QUESTION OF 218 BC: A FORENSIC EXERCISE IMPORTANCE TO HISTORICAL ARCHAEOLOGY 61

Mahaney, W. C., Allen, C. C. R., Pentlavalli, P., Dirszowsky, O., Tricart, P., Keiser, L., Somelar, P., Kelleher, B., Murphy, B., Costa,, P. J. M., and Julig, P. (2014) Polybius‘s ‗previous landslide‘: proof that Han- nibal‘s invasion route crossed the col de la Traversette, Journal of Mediterranean Archaeology and Ar- chaeometry, Vol. 14 (2), pp. 1–20. Mahaney, W. C., Somelar, P., Dirszowsky, R. W., Kelleher, B., Pentlavalli, P., McLaughlin, S., Kulakova, A., Jordan, S., Pulleyblank, C., West, A., Allen, C.C.R. (2016a) A microbial link to weathering of postglacial rocks and sediments, Mt. Viso area, Western Alps demonstrated through analysis of a soil/paleosol bio/chronosequence. Journal of Geology, Vol. 124 (2), pp. 149-169. Mahaney, W.C., Hancock, R.G.V., Somelar, P., Milan, A.( 2016b) Iron and Al Soil/Paleosol Extractions as Age/Environment Indicators: some examples from a catchment in Southern Ontario, Canada. Ge- omorphology, Vol. 270, pp. 159-171. Mahaney, W. C., Allen, C.C.R., Pentlavalli, P., Kulakova, A., Young, J. M., Dirszowsky, R.W., West, A., Kelleher, B., Jordan, S., Pulleyblank, C., O‘Reilly, S., Murphy, B. T., Lasberg, K., Somelar, P., Gar- neau, M., Finkelstein, S. A., Sobol, M. K., Kalm, V., Costa, P. J. M., Hancock, R. G. V., Hart, K. M., Tricart, P., Barendregt, R. W., Bunch, T. E., Milner, M. W. (2017a) Biostratigraphic Evidence relating to the Age-Old Question of Hannibal‘s Invasion of Italy: I, History and Geological Reconstruction, Archaeometry, Vol. 59, pp. 164-178. Mahaney, W. C., Allen, C. C. R., Pentlavalli, P., Kulakova, A., Young, J. M., Dirszowsky, R. W., West, A., Kelleher, B., Jordan, S., Pulleyblank, C., O‘Reilly, S., Murphy, B. T., Lasberg, K., Somelar, P., Gar- neau, M., Finkelstein, S. A., Sobol, M. K., Kalm, V., Costa, P. J. M., Hancock, R. G. V., Hart, K. M., Tricart, P., Barendregt, R. W., Bunch, T. E., and Milner, M. W. (2017b) Biostratigraphic evidence relating to the age-old question of Hannibal‘s invasion of Italy, II: chemical biomarkers and microbial signatures, Archaeometry Vol. 59, pp. 179-190. Mahaney, W.C., Somelar, P., Pulleyblank, C., Tricart, P., West, A., Young, J., and Allen, C.C.R. (2017c) Notes on magnetic susceptibility in the Guil valley alluvial mire correlated with the Punic Invasion of Ita- lia in 218 BC. Mediterranean Journal of Archaeometry and Archaeology, Vol. 17, (1), pp. 23-35. Mahaney, W.C., Somelar, P., West, A., Krinsley, D., Allen, C. C. R., Pentlavalli, P., Young, J. M., Dohm, J.M., LeCompte, M., Kelleher, B., Jordan, S., Pulleyblank, C., Dirszowsky, R., Costa, P., (2017d) Evidence for cosmic airburst/impact in the Western Alps archived in Late Glacial Paleosols. Quaternary In- ternational, Vol. 438, no. B, pp. 68-80. Mahaney, W.C., Peeter Somelar, Allen West, Randy Dirszowsky, Christopher C.R. Allen, Tarmo Remmel (2018a) Reconnaissance of the Hannibalic Route in the Upper Po Valley. Italy: Correlation with Bi- ostratigraphic Historical Archaeological Evidence in the Upper Guil Valley of France, Archaeometry, doi.org/10.1111/arcm.12405. Mahaney, W.C., West, Allen, Milan, Alison, Krinsley, David, Somelar, Peeter, Schwartz, Stephane, Milner, Michael W., Allen, Christopher C.R.(2018b). Cosmic impact/airburst on deposits/soils in the Western Alps of the Mt. Viso area, France, Quaternaria, Vol. 35 ( 1) 3–23 . Mahaney, W.C., Krinsley, D.H., Milner, Michael W. (2018c) Did the black mat/airburst reach the Antarctic: Evidence from New Mountain near the Taylor Glacier in the Dry Valleys. Journal of Geology, Vol. 126 (3), pp. 285-305. Miyamoto, H., Yano, H., Scheeres, D. J., Abe, S., Barnaouin-Jha, O., Cheng, A. F., Demura, H., Gaskell, R. W., Hirata, N., Ishiguro, M., Michikami, T., Nakamura, A. M., Hakamura, R., Saito, J., and Sasaki, S. (2007) Regolith migration and sorting on asteroid Itokawa, Science, Vol. 316, pp. 1011–1014. Millar, C. E., Turk, L. M., and Foth, H. D. (1966) Fundamentals of Soil Science, NY, 4th ed, Wiley. Mulley, Clare (2012) The Spy Who Loved, St. N.Y., Martins Griffin, NY, pp. 426. Neumann, J. (1992). Climatic conditions in the Alps in the years about the year of Hannibal‘s crossing (218 BC). Climatic Change, vol.22, pp.139–150. Oyama, M., and Takehara, H.(1970), Standard soil color charts, Japan Research Council for Agriculture, Forest- ry and Fisheries, Tokyo, Japan. Paton, W.R. (1922) Polybius, The Histories, Books 3-4, Cambridge, Ma., Harvard University Press, 568 pp. Polybius (1979) The rise of the Roman Republic, trans. I. Scott-Kilvert, pp. 215–29, London, Penguin. Prevas, J. (1998) Hannibal Crosses the Alps. Rockville Centre, NY Sarpedon, , 232 pp. Proctor, D. (1971) Hannibal’s march in history, Oxford, Oxford University Press. Reimer, P. J., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., Buck, C. E., Cheng, H., Edwards, R. L., Friedrich, M., Grootes, P. M., Guilderson, T. P., Haflidason, H., Hajdas, I., Hatté, C., Heaton, T. J., Hoffman, D., L., Hogg, A. G., Hughen, K. A., Kaiser, K. F., Kromer, B., Manning, S.

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62 62 W.C. MAHANEY

W., Niu, M., Reimer, R. W., Richards, D. A., Scott, E. M., Southon, J. R., Staff, R. A., Turney, C. S. M., and Van Der Plicht, J. (2013) IntCal13 and Marine13 radiocarbon age calibration curves 0– 50,000 years cal BP, Radiocarbon, Vol. 55 (4), pp. 1869–87. Romano, Marco, Palombo, Rita (2017) When legend, history and science rhyme: Hannibal‘s war elephants as an explanation to large vertebrate skeletons found in Italy, Historical Biology, DOI: 10.1080/08912963.2017.1287178 To link to this article: ttp://dx.doi.org/10.1080/08912963.2017.1287178. Seibert, J. (1993) Hannibal, pp. 75–134, Darmstadt, Wissenschaftliche Buchgesellschaft,. Sodhi, R., Mahaney, W. C., and Milner, M. W. (2006) ToF-SIMS applied to historical archaeology in the Alps, Applied Surface Science, Vol. 252, pp. 7140–3. Soil Survey Staff (1999) Soil taxonomy, USDA Agricultural Handbook No. 436, Government Printing Office, Washington, DC. Strabo, trans. H.L, Jones (1999) Strabo: Geography Vol II. Harvard University Press, Copyright 1923. Strabo‘s map of the Mediterranean Basin. Tricart, P., Schwartz, S., Lardeaux, J. -M., Thouvenot, F., and Du Chaffaut, S. A. (2003) Aiguilles-Col Saint- Martin, Carte Géologique de la France, 1:50000. Walbank, F. W. (1956) Some reflections on Hannibal‘s Pass, Journal of Roman Studies, Vol. 46, pp. 37–45. Walbank, F. W. (1990) Polybius, Berkeley, CA, University of California Press,. Wilkinson, H. S. (1911) Hannibal’s march through the Alps, Oxford, The Clarendon Press. Wolbach, W.S., Ballard, J.P., Mayewski, P.A., Parnell, A.C., Cahill, N., Adedeji, V., Bunch, T.E., Domínguez- Vázquez, G., Erlandson, J.M., Firestone, R.B., French, T.A., Howard, G., Israde-Alcántara, I., John- son, J.R., Kimbel, D., Kinzie, C.R., Kurbatov, A., Kletetschka, G., LeCompte, M.A., Mahaney, W.C., Melott, A.L., Mitra, S., Maiorana-Boutilier, A., Moore, C.R., Napier, W.M., Parlier, J., Tankersley, K.B., Thomas, B.C., Wittke, J.C., West, A., Kennett, J.P. (2018a) Extraordinary Biomass-Burning Epi- sode and Impact Winter Triggered by the Younger Dryas Cosmic Impact~12,800 Years Ago. 1. Ice Cores and Glaciers. Journal of Geology, Vol. 126 (2), pp. 165–84. Wolbach, W.S., Ballard, J.P., Mayewski, P.A., Parnell, A.C., Cahill, N., Adedeji, V., Bunch, T.E., Domínguez- Vázquez, G., Erlandson, J.M., Firestone, R.B., French, T.A., Howard, G., Israde-Alcántara, I., John- son, J.R., Kimbel, D., Kinzie, C.R., Kurbatov, A., Kletetschka, G., LeCompte, M.A., Mahaney, W.C., Melott, A.L., Mitra, S., Maiorana-Boutilier, A., Moore, C.R., Napier, W.M., Parlier, J., Tankersley, K.B., Thomas, B.C., Wittke, J.C., West, A., Kennett, J.P. (2018b) Extraordinary Biomass-Burning Ep- isode and Impact Winter Triggered by the Younger Dryas Cosmic Impact ∼12,800 Years Ago. 2. Lake, Marine, and Terrestrial Sediments, Journal of Geology, Vol. 126 (2), pp. 185-205.

Mediterranean Archaeology and Archaeometry, Vol. 18, No 3, (2018), pp. 45-62